JP5216977B2 - Fruit discrimination structure - Google Patents

Description

  The present invention relates to a fruit discrimination structure for discriminating the quality of fruit.

  As a method for finely selecting the green soybean cake, there is a method in which an image of the green soybean cake is acquired by a CCD camera, and an outer shape and a discolored portion of the green soybean cake are extracted by an image processing unit (see, for example, Patent Document 1).

Here, it is preferable that the quality of the soy bean pod can be properly discriminated in such a method for selecting soy bean pods.
JP 2005-279524 A

  An object of the present invention is to obtain a fruit determination structure that can appropriately determine the quality of fruit in consideration of the above facts.

The fruit discriminating structure according to claim 1 is an image pickup means for picking up an image of a fruit by at least one of transmitted light transmitted through the fruit and reflected light reflected by the fruit; and a longitudinal end portion of the fruit image as a fruit. Discriminating means for discriminating the tear state of the longitudinal end portion of the fruit based on whether or not there is a pixel in which the fruit is not imaged between pixels in which the fruit is imaged on a pass line passing in the width direction of the image of It has.

The fruit discriminating structure according to claim 2 passes through the center of gravity based on the area of the fruit image and the imaging means for picking up the fruit image by at least one of the transmitted light transmitted through the fruit and the reflected light reflected by the fruit. The shape of the fruit is determined based on the ratio between the length of the fruit image along the longitudinal line in the longitudinal direction of the fruit image and the length of the fruit image along the width line in the width direction of the fruit image passing through the center of gravity. Discriminating means for discriminating.

According to a third aspect of the present invention , the fruit discrimination structure has an imaging means for picking up an image of the fruit by at least one of the transmitted light transmitted through the fruit and the reflected light reflected by the fruit, and the center of gravity based on the area of the fruit image is the fruit. Discriminating means for discriminating the shape of the fruit based on whether or not it exists within the range of the image.

The fruit discrimination structure according to claim 4 is the fruit discrimination structure according to any one of claims 1 to 3 , further comprising binarizing means for binarizing the fruit image. It is a feature.

The fruit discrimination structure according to claim 5 is characterized in that, in the fruit discrimination structure according to any one of claims 1 to 4 , the discrimination means discriminates the grade of the fruit.

In the fruit discrimination structure according to the first aspect, the imaging means captures an image of the fruit by at least one of the transmitted light that has passed through the fruit and the reflected light that has been reflected by the fruit. Further, whether or not there is a pixel in which the fruit is not captured between the pixels in which the fruit is captured on the pass line passing through the longitudinal end of the fruit image in the width direction of the fruit image. Based on this, the tearing state of the longitudinal end of the fruit is determined. For this reason, the quality of the fruit can be determined appropriately.

In the fruit discrimination structure according to claim 2 , the imaging means captures an image of the fruit by at least one of the transmitted light that has passed through the fruit and the reflected light that has been reflected by the fruit. Further, the discriminating means adds the length of the fruit image along the longitudinal line of the fruit image passing through the center of gravity based on the area of the fruit image and the width line of the width of the fruit image passing through the center of gravity. The shape of the fruit is determined based on the ratio of the length of the image along the fruit. For this reason, the quality of the fruit can be determined appropriately.

In the fruit discrimination structure according to the third aspect, the imaging means captures an image of the fruit by at least one of the transmitted light transmitted through the fruit and the reflected light reflected by the fruit. Further, the determining means determines the shape of the fruit based on whether or not the center of gravity based on the area of the fruit image exists within the range of the fruit image. For this reason, the quality of the fruit can be determined appropriately.

In the fruit discrimination structure according to the fourth aspect , the binarization means binarizes the fruit image. For this reason, the discrimination means can easily discriminate the quality of the fruit.

In the fruit discrimination structure according to the fifth aspect, the discrimination means discriminates the grade of the fruit. For this reason, the quality of the fruit can be determined clearly.

  FIG. 8 is a schematic view of the fruit discrimination structure 10 according to the embodiment of the present invention viewed from the right side. In the drawing, the front of the fruit discrimination structure 10 is indicated by an arrow FR, and the upper part of the fruit discrimination structure 10 is indicated by an arrow UP.

  The fruit discriminating structure 10 according to the present embodiment includes a belt conveyor 12 as a conveying means. In the belt conveyor 12, an endless belt 16 is wound around a pair of rollers 14. When the belt conveyor 12 is driven, the belt 16 is rotated and the upper portion of the belt 16 is moved forward. At the same time, the lower portion of the belt 16 is moved rearward. Further, the belt 16 is formed into a net-like net, for example, so that light can be transmitted.

  A pair of long plate-like partition plates 18 are fixed immediately above the belt 16, and the pair of partition plates 18 are each extended in the front-rear direction to form a conveyance path 20 therebetween. For this reason, by placing the fruit 22 (in the present embodiment, the fruit of edamame) on the conveyance path 20 on the upper portion of the belt 16, the belt 16 is placed with the fruit 22 oriented in the longitudinal direction. Is conveyed forward (conveying direction X). In addition, the fruit 22 placed on the conveyance path 20 is the one after having been detached from the branch by a dehuller (not shown) or by manual work, and the fruit 22 is usually in the basket 24. A predetermined number of beans 26 (grains) are accommodated (see FIG. 1, FIG. 1 exemplifies the fruit fruit 22 in which two beans 26 are accommodated in the basket 24).

  A camera 28 (for example, a CCD camera) as an imaging unit is provided above the belt 16, and when the fruit 22 conveyed forward by the belt 16 is disposed below the camera 28, the camera 28 is The fruit 22 can be imaged from above.

  A rectangular parallelepiped box-shaped camera booth 30 as a dark room forming means is provided above the belt 16, and the lower surface of the camera booth 30 is opened and disposed immediately above the upper portion of the belt 16. The lower end portions of the front wall and the rear wall of the camera booth 30 are opened in a rectangular shape in the range between the pair of partition plates 18, and thereby, into the camera booth 30 of the fruit 22 conveyed forward by the belt 16. In and out of the camera booth 30 are not hindered. Inside the camera booth 30, the lens portion 28A of the camera 28 is inserted from the upper wall of the camera booth 30, and the camera booth 30 forms a dark room 32 therein, and the image 22 captured by the camera 28 is captured. The area is a dark room 32.

  Below the belt 16, a transmitted light irradiation device 34 as a transmitted light irradiation means is provided at a position below the camera 28, and the transmitted light irradiation device 34 moves the belt 16 to the fruit 22 imaged by the camera 28. Thus, it is possible to irradiate light from below with equal intensity.

  In the camera booth 30, reflected light lamps 36 are provided as reflected light irradiating means on the front side and the rear side of the camera 28, and the pair of reflected light lamps 36 are provided in front of the fruit 22 captured by the camera 28. By irradiating light from diagonally upward and diagonally upward, light can be irradiated to the entire upper side surface (one side surface) of the fruit 22.

  The camera 28 is provided with a control device 38 as binarization means and discrimination means, and the control device 38 binarizes the image of the fruit 22 imaged by the camera 28 and also picks up the imaged fruit. Based on the 22 captured images or the binarized image of the binarized fruit 22, the quality of the fruit 22 can be determined.

  Further, at present, the fruit 22 is discriminated as an excellent product, a non-defective product, and a defective product according to the quality standard shown in FIG. It is distinguished by the grade.

  Next, the operation of the present embodiment will be described.

  In the fruit discrimination structure 10 having the above-described configuration, the fruit 22 is placed on the conveyance path 20 between the pair of partition plates 18 on the upper portion of the belt 16 in a state where the belt conveyor 12 is driven. Are conveyed forward (conveying direction X) by the belt 16 with the longitudinal direction thereof directed in the front-rear direction. Further, when the fruit 22 conveyed forward by the belt 16 is arranged below the camera 28 in the camera booth 30 (in the dark room 32), the transmitted light irradiation device 34 is lowered to the fruit 22 via the belt 16. The camera 28 images the fruits 22 from above with the transmitted light that has passed through the fruits 22 by irradiating light with equal intensity.

  The captured image of the fruit 22 shown in FIG. 1A captured by the camera 28 is binarized by the control device 38 with a predetermined threshold value, and the binarization level 1 at which the fruit 22 (the entire fruit 24) is imaged. Other than the binarized image (see FIG. 1 (B)), the binarized level 2 binarized image (see FIG. 1 (C)) in which the beans 26 of the fruits 22 are imaged, and the beans 26 of the fruits 22 The binarized level 2 binarized image (see FIG. 1 (D); the binarized image of FIG. 1 (C) is displayed in reverse) is created.

  The control device 38 calculates the ratio of the total number of pixels of the beans 26 (corresponding to the total area of the beans 26) to the number of pixels of the fruits 22 (corresponding to the area of the fruits 22) in the binarized image of the fruits 22. Thus, the state of accommodation of the beans 26 in the cocoon 24 in the coconut fruit 22 (the actual state, the maturity) is determined. In addition, the control device 38 determines whether the fruit 22 is contained in the cocoon 24 based on the magnitude relationship between the ratio of the number of pixels of the beans 26 to the number of pixels of the fruit 22 and a predetermined number of thresholds. It is determined whether the product is a good product or a defective product (immature product or over-mature product), or whether the fruit 22 is an excellent product, a good product, or a defective product. It is determined which of the grades. Thereby, the quality of the fruit fruit 22 about the accommodation state of the beans 26 in the basket 24 can be determined appropriately and clearly. In addition, since the fruit fruit 22 is a fruit of a green soybean, even if the fruit fruit 22 is an overripe product, the beans 26 of the fruit fruit 22 are separately used as soybeans.

  In addition, the control device 38 determines the size of the fruit 22 by calculating the number of pixels of the fruit 22 in the captured image of the fruit 22 or the binarized image of the fruit 22 (binarization level 1 image). In addition, the control device 38 determines whether the fruit 22 is a non-defective product or a defective product with respect to the size of the fruit 22 based on the magnitude relationship between the number of pixels of the fruit 22 and a predetermined number of thresholds. It is determined whether the fruit 22 is an excellent product, a non-defective product, or a defective product, thereby determining which of the plurality of grades the fruit 22 is. Thereby, the quality of the fruit 22 about a magnitude | size can be discriminate | determined appropriately and clearly.

  Further, as shown in FIGS. 2 and 3A and 3B, the control device 38 selects the fruit 22 in the picked-up image of the fruit 22 or the binarized image of the fruit 22 (binarization level 1 image). A predetermined distance L (distance equal to or shorter than the length of the fruit stem 24A in the longitudinal direction of the fruit 22) from both ends in the longitudinal direction (the front-rear direction along the transport direction X by the belt conveyor 12). On the pair of horizontal straight lines A and B (passing lines) passing through the portion in the width direction of the fruit 22 (left and right direction perpendicular to the longitudinal direction of the fruit 22), the pixel on which the fruit 22 is imaged (the pixel of the fruit 22 In this embodiment, it is determined whether there is a pixel (a non-fruit 22 pixel, which is a relatively bright pixel in this embodiment) where the fruit 22 is not captured in a relatively dark pixel) The torn state of the fruit fruit 22 (especially the fruit stem 24A Torn state) to determine. That is, when there is a non-fruit 22 pixel among the pixels of the fruit 22 on at least one of the pair of horizontal straight lines A and B, the control device 38 determines that the fruit 22 is torn and has a defective product. If there is no non-fruit 22 pixel in the fruit 22 pixel on both the pair of horizontal straight lines A and B, the fruit 22 is not torn and the fruit 22 is determined to be non-defective. It is discriminate | determined. Thereby, the quality of the fruit 22 about a torn state can be determined appropriately and clearly. 3A and 3B show pixel waveforms of captured images on the horizontal straight lines A and B. FIG.

  In place of this, the control device 38 moves from the gravity center G based on the area of the fruit 22 in the longitudinal direction of the fruit 22 in the captured image of the fruit 22 or the binarized image of the fruit 22 (binarization level 1 image). On the far side horizontal straight line A, it is also possible to determine whether or not there is a non-fruit 22 pixel in the fruit 22 pixel, and to determine the tear state of the fruit stem 24 A of the fruit 22. That is, since the fruit fruit 22 is always on the side where the fruit stem 24A is farther from the center of gravity G than the opposite end, the control device 38 determines whether or not the non-fruit 22 pixel exists in the fruit 22 pixel on the horizontal straight line A. If there is a tear in the fruit stem 24A, it is determined as a defective product, and if it is not present, it is determined as a good product because there is no tear in the fruit stem 24A. Determine which of the grades. Thereby, the quality of the fruit fruit 22 about the tear state of the fruit stem 24A can be determined appropriately, clearly and simply.

  Further, as shown in FIGS. 4A to 4C, the control device 38 sets the area of the fruit 22 in the captured image of the fruit 22 or the binarized image of the fruit 22 (binarization level 1 image). Along the long axis straight line S (longitudinal line) in the longitudinal direction of the fruit 22 passing through the center of gravity G and along the short-axis straight line T (width line) in the width direction of the fruit 22 passing through the center of gravity G. The ratio of the fruit 22 is calculated to determine the shape of the fruit 22. For example, as shown in FIG. 4A, the control device 38 has a ratio of the length of the fruit 22 along the short-axis straight line T to the length of the fruit 22 along the long-axis straight line S smaller than the first threshold (less than 1). If it is less than 0), it is determined that the fruit 22 is a fine product having a thin shape, and the short axis straight line T with respect to the length of the fruit 22 along the long straight line S as shown in FIG. When the ratio of the length of the fruit 22 along the line is equal to or greater than the first threshold and less than the second threshold (a value smaller than 1 and close to 1), the fruit 22 is determined to be an excellent product having a normal shape, and FIG. ), When the ratio of the length of the fruit 22 along the short-axis straight line T to the length of the fruit 22 along the long-axis straight line S is equal to or greater than the second threshold value, the fruit 22 has an abnormal shape. (For example, one bean 26 is determined), and it is determined which of a plurality of grades the fruit fruit 22 is. Thereby, the quality of the fruit 22 about a shape can be discriminate | determined appropriately and clearly. Note that the control device 38 may simply determine that the fruits 22 determined as good and excellent in FIGS. 4A and 4B are good.

  Further, as shown in FIGS. 5A and 5B, the control device 38 sets the area of the fruit 22 in the captured image of the fruit 22 or the binarized image of the fruit 22 (binarization level 1 image). It is determined whether or not the based center of gravity G exists within the range of the fruit 22, and the shape of the fruit 22 is determined. That is, when the center of gravity G exists within the range of the fruit 22 as shown in FIG. 5A, the control device 38 determines that the fruit 22 is a good product having a normal shape, and the center of gravity as shown in FIG. If G does not exist within the range of the fruit 22, the fruit 22 is determined as a defective product having an abnormal shape (for example, a curved shape), and it is determined which of the plurality of grades the fruit 22 is. Thereby, the quality of the fruit 22 about a shape can be discriminate | determined more appropriately. FIGS. 4A and 4B and FIGS. 5A and 5B illustrate the fruit 22 in which three beans 26 are accommodated in the basket 24.

  Further, in the fruit discrimination structure 10 according to the present embodiment, when the fruit 22 conveyed forward by the belt 16 is disposed below the camera 28 in the camera booth 30 (in the dark room 32), a pair of reflected lights. When the lamp 36 irradiates the fruit 22 with light (visible region (RGB) or near-infrared region (IR)) from the front obliquely upper side and the rear oblique upper side, the reflected light reflected by the fruit 22 causes the camera 28 to move upward. Then, the fruit 22 is imaged.

  6A to 6E captured by the camera 28 are spectrally extracted by the spectral filter by the control device 38 and binarized with a predetermined threshold value to obtain the result 22 ( A binarized level 1 binarized image (see FIGS. 7 (A-1) to (E-1)) and the damaged portion 22A of the fruit 22 (莢 24) are imaged. A binarized level 2 binary image (see FIGS. 7A-2 to 7E-2) and a normal portion 22B of the fruit 22 (莢 24). The converted image (see FIGS. 7A-3 to E-3; in which FIGS. 7A-2 to E-2 are highlighted) is created.

  In the binarized image of the fruit 22, the control device 38 is a ratio of the number of pixels of the damaged part 22A of the fruit 22 (corresponding to the area of the damaged part 22A) to the number of pixels of the fruit 22 (corresponding to the area of the fruit 22). Is calculated to determine whether the fruit 22 is damaged. In addition, the control device 38 determines whether the fruit 22 is in the damaged state of the fruit 22 based on the magnitude relationship between the ratio of the number of pixels of the damaged portion 22A of the fruit 22 to the number of pixels of the fruit 22 and a predetermined threshold value. Whether it is a non-defective product (for example, the ratio is less than 8%) or a defective product (for example, the ratio is 8% or more), or the fruit 22 is a superior product (for example, the ratio is less than 1%) and a non-defective product (for example, the ratio is less than 1%). For example, it is determined whether the ratio is 1% or more and less than 8%) or a defective product (for example, the ratio is 8% or more), and it is determined whether the fruit 22 is a plurality of grades. . Thereby, the quality of the fruit 22 about a damage state can be discriminate | determined appropriately and clearly.

  Moreover, the normal part 22B of the fruit fruit 22 is green. On the other hand, the damaged portion 22A of the fruit fruit 22 is caused by pest damage or damage such as when the fruit fruit 22 is detached from the branch. Normally, the damaged part 22A of the fruit fruit 22 is rusted (compressed) or compressed (bean paste). ), Discoloration (in particular discoloration to brown or black), and scratches. Here, as shown in FIG. 9, even if the damaged portion 22A of the fruit 22 is rust (cracking), compressed (carried), discolored, or scratched, the spectrum of the normal portion 22B and the damaged portion 22A of the fruit 22 is separated. The difference in reflectance increases in the wavelength region of 540 nm or 730 nm. For this reason, in the binarized image of the fruit 22 by the binarization process in which the reflected light reflected by irradiating the fruit 22 with light having a wavelength range of 540 nm or 730 nm from the pair of reflected light lamps 36 is spectrally extracted. The normal portion 22B and the damaged portion 22A can be distinguished well. Thereby, the quality of the fruit 22 about a damage state can be discriminate | determined more appropriately.

  As described above, the quality of the fruit 22 can be mechanically determined, and the selection based on the quality of the fruit 22 can be automated.

  In the present embodiment, in the captured image of the fruit 22 captured by the transmitted light transmitted through the fruit 22 or the binarized image of the captured image, the size of the fruit 22 and the torn state of the fruit 22 (fruit (Including the tear state of the infarction 24A) and the shape of the fruit 22, but in the captured image of the fruit 22 captured by the reflected light reflected by the fruit 22, or the binarized image of the captured image The size of the fruits 22, the torn state of the fruits 22 (including the torn state of the fruit stem 24 </ b> A), and the shape of the fruits 22 may be determined.

  In the present embodiment, the ratio of the number of pixels of the beans 26 to the number of pixels of the fruits 22 is calculated to determine the accommodation state of the beans 26 in the fruits 24. However, the number of pixels of the fruits 22 and the beans 26 are determined. And the number of pixels other than the beans 26 (equal to the number of pixels of the fruit 22 minus the number of pixels of the beans 26) are determined to determine the accommodation state of the beans 26 in the basket 24 Any configuration may be used.

  Furthermore, in the present embodiment, the ratio of the number of pixels of the damaged portion 22A of the fruit 22 to the number of pixels of the fruit 22 is calculated to determine the damage state of the fruit 22. However, the number of pixels of the fruit 22 and the fruit 22 are determined. Damage of the fruit 22 based on at least two ratios of the number of pixels of the damaged part 22A and the number of pixels of the normal part 22B of the fruit 22 (equal to the number of pixels of the fruit 22 minus the number of pixels of the damaged part 22A) Any configuration that determines the state may be used.

  In the present embodiment, when the quality of the fruit 22 is determined based on the image of the fruit 22 captured by the reflected light reflected by the fruit 22, the fruit is based on the image of the upper side surface (one side surface) of the fruit 22. However, since the upper side (one side) and the lower side (other side) of the fruit 22 are generally in the same state, the image of the one side and the other side of the fruit 22 is not necessarily the same. It is not necessary to determine the quality of the fruit 22 based on the above, and even if the quality of the fruit 22 is determined based on the image of only one side of the fruit 22, the quality of the fruit 22 can be appropriately determined.

  Further, in the present embodiment, green soybean fruits are used as the fruits 22, but kidney beans, peas, or broad beans fruits may be used as the fruits 22.

(A) is a figure for the bean accommodation state determination of the fruit which shows the picked-up image of the fruit which was imaged with the transmitted light which permeate | transmitted the fruit in the fruit determination structure which concerns on embodiment of this invention, (B) FIG. 5B is a diagram for determining the bean accommodation state of a fruit and showing a binarized image of the fruit and fruit picked up by the transmitted light that has been transmitted through the fruit in the fruit determination structure according to the embodiment of the present invention. FIG. 4 is a diagram for determining a bean accommodation state of a fruit and showing a binarized image of the fruit and beans imaged by transmitted light that has been transmitted through the fruit in the fruit determination structure according to the embodiment of the present invention; It is a figure for the bean accommodation state determination of the fruit which shows the binarized image of parts other than the bean of the fruit which were imaged with the transmitted light which permeate | transmitted the fruit in the fruit discrimination structure which concerns on embodiment of this invention. It is a figure for the torn state determination of the fruit which shows the image of the fruit in the fruit discrimination structure which concerns on embodiment of this invention. (A) is a waveform showing the brightness of the pixel on the horizontal line A in FIG. 2, and (B) is a waveform showing the brightness of the pixel on the horizontal line B in FIG. (A) thru | or (C) is a figure for thickness shape determination of the fruit which shows the image of the fruit in the fruit discrimination structure which concerns on embodiment of this invention. (A) And (B) is a figure for curve shape discrimination | determination of the fruit which shows the image of the fruit in the fruit discrimination structure which concerns on embodiment of this invention. (A) thru | or (E) is a figure for the damage state determination of the fruit which shows the picked-up image of the fruit imaged with the reflected light reflected in the fruit in the fruit discrimination structure which concerns on embodiment of this invention. (A-1) thru | or (E-1) are the damage state of the fruit which shows the binarized image of the fruit of a fruit imaged with the reflected light reflected by the fruit in the fruit discrimination structure which concerns on embodiment of this invention It is a figure for discrimination | determination, (A-2) thru | or (E-2) are the binary values of the damaged part of the fruit image imaged by the reflected light reflected by the fruit fruit in the fruit judgment structure which concerns on embodiment of this invention. FIG. 4A is a diagram for determining a damage state of a fruit and showing (A-3) to (E-3) captured by reflected light reflected on the fruit in the fruit determination structure according to the embodiment of the present invention; It is a figure for the damage state discrimination | determination of the fruit which shows the binarized image of the normal part of the fruit. It is the schematic diagram seen from the right side which shows the fruit discrimination structure which concerns on embodiment of this invention. It is a graph which shows the relationship between the wavelength of the light irradiated to the fruit in the fruit discrimination structure which concerns on embodiment of this invention, and the difference of the spectral reflectance of the reflected light reflected in the normal part and damaged part of a fruit . It is a table | surface which shows the present fruit quality standard.

Explanation of symbols

10 Fruit recognition structure 22 Fruit 22A Damaged part 24 Coffee 26 Bean 28 Camera (imaging means)
38 Control device (binarization means, discrimination means)
A Horizontal straight line (passing line)
B Horizontal straight line (passing line)
G Center of gravity S Long axis straight line (longitudinal line)
T short axis straight line (width line)

Claims (5)

Imaging means for capturing an image of the fruit by at least one of the transmitted light transmitted through the fruit and the reflected light reflected by the fruit;
The longitudinal end of the fruit based on whether or not there is a pixel in which the fruit is not captured between the pixels in which the fruit is captured on a pass line passing through the longitudinal end of the fruit image in the width direction of the fruit image A discriminating means for discriminating the tearing state of
Fruit discriminating structure with Imaging means for capturing an image of the fruit by at least one of the transmitted light transmitted through the fruit and the reflected light reflected by the fruit;
The length of the fruit image along the longitudinal line of the fruit image passing through the center of gravity based on the area of the fruit image and the width of the fruit image along the width line of the fruit image passing through the center of gravity. Discriminating means for discriminating the shape of the fruit based on the ratio with the length;
Fruit discriminating structure with Imaging means for capturing an image of the fruit by at least one of the transmitted light transmitted through the fruit and the reflected light reflected by the fruit;
Discrimination means for discriminating the shape of the fruit based on whether or not the center of gravity based on the area of the fruit image exists within the range of the fruit image;
Fruit discriminating structure with The fruit discrimination structure according to any one of claims 1 to 3, further comprising binarization means for binarizing the fruit image. The fruit discrimination structure according to any one of claims 1 to 4, wherein the discrimination means discriminates a grade of fruit.